the economic roots of external efficacy: assessing the … · the economic roots of external...

The Economic Roots of External Efficacy:Assessing the Relationship between ExternalPolitical Efficacy and Income Inequality

MIKEL NORRIS Coastal Carolina University

Introduction

This paper addresses and connects two important concepts in Americanpolitics: external political efficacy and income inequality. External politicalefficacy—the belief that government officials and institutions are respon-sive to the demands of its citizens—has steadily declined since the1960s. External political efficacy’s decline is often viewed as a symptomof greater problems in the American political system. It is most notablyused as an explanatory factor in understanding different forms of politicalparticipation (Bowler and Donovan, 2002; Chamberlain, 2013; Finkel,1985; Platt, 2008; Pollock, 1983; Valentino et al., 2009) and voterturnout (Abramson and Aldrich, 1982; Mangum, 2003). However, scholarshave been unsuccessful in explaining the reasons why external political effi-cacy in the American electorate has steadily declined. This is unfortunate.Since the United States is a democratic society, it is vitally important thatthe voices of all political participants are heard, and studying the causesof changes in external political efficacy is essential for understandinghow democratic political systems operate (Chamberlain, 2012).

Income inequality is also an important area of study in American poli-tics. Political decisions have distributional consequences, and one of themost important distributional outcomes of the political process is determin-ing how the economic pie will be divided (Kelly, 2005). Income inequalityin the United States has risen rapidly over the last thirty years and a lot of

Mikel Norris, Department of Politics and Geography, Coastal Carolina University, POBox 261954, Conway SC 29528-6054, Email: [email protected]

Canadian Journal of Political Science / Revue canadienne de science politiquePage 1 of 23 doi:10.1017/S0008423915001080

© 2016 Canadian Political Science Association (l’Association canadienne de science politique)and/et la Société québécoise de science politique

mailto:[email protected]

research is devoted to understanding the effects of growing income inequa-lity on the American political system (Bartels, 2008; Gilens, 2005, 2012;Hacker and Pierson, 2010; McCarty et al., 2006). While there are someconflicting findings (see, for example, Bennett, 2006; Soroka andWlezien, 2008), most analyses point to increasing income inequality nega-tively affecting the American political system. Specifically, researchdemonstrates that income inequality may shape the political agenda bycrowding out policy alternatives important to the middle class and poorin favour of policies of importance to wealthier citizens. The crowdingout of policies important to the middle class and the poor in turn leads todecreasing political participation and, ultimately, lower voter turnout(Schattschneider, 1960; Solt, 2010).

Research on income inequality implies that increasing income inequa-lity leads to lower levels of external political efficacy in the electorate.However, this has never been tested empirically.1 This paper attempts tofill this void in the research by examining income inequality’s effect onexternal political efficacy. Using a multilevel dataset that combines indivi-dual attributes with state and yearly information on electoral and economiccontexts, my results provide evidence that increasing income inequality inthe American states has important negative consequences for levels ofexternal political efficacy in the American electorate.

External Political Efficacy: Definition, Measurement and Decline

American politics scholars have studied political efficacy for decades.Initially, scholars defined political efficacy as the feeling that one’sactions can and do have an impact on the political process (Craig et al.,1990). Scholars have since broken political efficacy into two components:internal political efficacy and external political efficacy. Internal politicalefficacy is defined as one’s belief in his or her ability to understand andeffectively participate in politics. External political efficacy, on the otherhand, measures one’s belief that political leaders and government institu-tions are responsive to the demands of its citizens (Balch, 1974).

Although these two concepts are distinct, scholars have encounteredproblems measuring both concepts and at times have confused the two.There is debate about how to correctly measure external political efficacy,and scholars have attempted to capture properties inherent in external poli-tical efficacy’s definition, including response variation across time and thedifferent effects local, state and national politics may have on external poli-tical efficacy (Chamberlain, 2012; Craig et al., 1990). However, scholarsagree that despite their flaws, two questions in the American NationalElection Studies tap external political efficacy’s true definition. The firstquestion asks respondents to respond to the statement: “People like me

2 MIKEL NORRIS

don’t have any say in what government does.” The second question asksrespondents to respond to the statement: “I don’t think public officialscare much what people like me think.” The answers to these questionsare then combined into an index that measures individual and aggregateexternal political efficacy levels (Chamberlain, 2012; Craig et al., 1990;Niemi et al., 1991).2

The interesting problem facing scholars is to explain why levels ofexternal political efficacy in the American electorate are steadily decliningand what effects this decline has on the political system. While levels ofinternal political efficacy have steadily increased over time, levels of exter-nal political efficacy have consistently declined since the 1960s. In aseminal study, Abramson and Aldrich (1982) credit the decline in externalpolitical efficacy—in conjunction with declining partisanship—as a keyreason for declines in national electoral participation. These two factorsmay have impacted voter turnout from the 1960s to the 1980s, but sincethen, partisanship and national electoral participation have risen (Bartels,2000). This has led some to think that Abramson and Aldrich’s claims ofexternal political efficacy’s effect on turnout may be overstated

Abstract. External political efficacy, the belief that government is responsive to the demands ofits citizens, has been declining in the United States since the 1960s. However, scholars do not yetfully understand the reasons for its decline. Nor have they found suitable explanations for why itfluctuates within the electorate. Drawing on the growing literature on the effects of income inequa-lity on public policy, I posit that increasing income inequality factors into the decline of externalpolitical efficacy. Using multilevel regression models accounting for individual and contextualfactors, I find increasing state-level income inequality has a substantial negative effect on externalpolitical efficacy. It is greater than most state and national-level economic measures or individual-level variables on external political efficacy. These results have important implications both forresearch on income inequality and political participation and also for research on income inequalityand distributional public policy.

Résumé. Aux États-Unis, l’efficacité politique extérieure—la conviction que le gouvernementrépond aux demandes de ses citoyens—est en constant déclin depuis les années soixante.Néanmoins, les chercheurs ne comprennent pas complètement les raisons de ce déclin. Ils n’ontpas trouvé non plus d’explications satisfaisantes à ces fluctuations parmi les électeurs. En faisantappel au volume croissant de la littérature au sujet des effets de l’inégalité de revenus sur la poli-tique publique, je pose comme principe que l’augmentation de l’inégalité des revenus contribue audéclin de l’efficacité politique extérieure. En utilisant les modèles de régression à plusieurs niveauxpour les facteurs individuels autant que contextuels, je trouve que l’augmentation de l’inégalité desrevenus au niveau des États a un effet négatif fondamental sur l’efficacité politique extérieure. Ceteffet a un impact sur l’efficacité politique extérieure plus important que celui de la plupart desmesures économiques au niveau national ou étatique, ou que celui des variables au niveau indivi-duel. Ces résultats ont des implications importantes sur la recherche sur l’inégalité des revenus et laparticipation politique, ainsi que sur la recherche sur l’inégalité des revenus et la politique publiquede distribution.

(Timpone, 1998). A possible explanation may be the concomitant rise andmoderating effect of internal political efficacy in the American electorate.Individuals with high internal political efficacy are likely to vote becausethey see themselves as capable participants in the political process. Yetthese same individuals may nevertheless have low external political effi-cacy because they believe the groups to which they belong are unrepresen-ted in the political system (Craig, 1980; Mangum, 2003; Michelson, 2000;Schur and Kruse, 2000; Teixeira, 1987).

Nor can changing political cultures account for external political effi-cacy’s decline. Chamberlain (2013) demonstrates that before 1980 indivi-duals embedded in individualistic, moralistic and traditionalistic politicalcultures had different levels of external political efficacy. However, diffe-rences among cultures disappeared beginning in 1980, and now levels ofexternal political efficacy in these cultures are similar and move intandem with one another over time. His analysis concludes that politicalculture no longer has a substantive effect on external political efficacy.

Declining political trust can also not be considered a culprit in externalpolitical efficacy’s decline. While these two concepts are related, they aredistinct from one another. External political efficacy measures a respon-dent’s belief that government is responding to citizens’ demands regardlessof the quality of its outputs. Political trust, on the other hand, is based on anormative belief about the quality of government outputs (Chamberlain,2012; Craig, 1980; Hetherington, 1998; Pollock, 1983). Although bothexternal political efficacy and political trust have declined over time, thecause of their simultaneous decline is still unclear, for different processesappear to be at work in the declines of each. For example, Anderson(2010) shows that fostering a sense of community raises external politicalefficacy and personal trust but not political trust. Likewise, Craig hypothe-sizes that external political efficacy is an important predictor of politicaltrust. When speaking of political mobilization, he states that, “external poli-tical efficacy may produce a behavioral response regardless of one’s fee-lings of trust or cynicism… In addition, cynicism will be a function ofspecific dissatisfactions other than external inefficacy” (Craig, 1980:206). Craig also states that, “the recent decline in external inefficacy mayhave been moderated by rising levels of education and internal efficacyamong the mass public… These factors are essentially unrelated to politicaltrust” (206). Although it has been posited that trust may affect external poli-tical efficacy, no empirical evidence of this relationship has been found(Chamberlain, 2012).

Suffice it to say that while the extant research on external political effi-cacy acknowledges its effect on electoral participation and political partici-pation more generally, it has not fully succeeded in finding the underlyingcauses as to why external political efficacy has declined and stayed at verylow levels since the 1960s. A simple and reasonable answer I propose as a

4 MIKEL NORRIS

solution to this puzzle is the following: low levels of external political effi-cacy in the United States can be explained through economic outcomes, andparticularly through growing income inequality. Several studies havelooked at economic indicators to measure changes in external political effi-cacy (Chamberlain, 2012; Lewis-Beck and Stegmaier, 2000). To date,however, none of the economic indicators used accounts for external poli-tical efficacy’s decline. In the next section, I will discuss the reasons whyincreasing income inequality in the United States may be linked to declininglevels of external political efficacy.

External Political Efficacy and the Economy: The Role of IncomeInequality

Although several analyses demonstrate that increasing income inequalitycauses predictable changes in public opinion (see, for example, Gilens,2009; Kelly and Enns, 2010; McCarty et al., 2006), none directly testwhether increasing income inequality in the United States may be linkedto declining levels of external political efficacy. However, several analysespresent findings that lead to the inference that there may be a connectionbetween the two concepts.

Research demonstrates that both rich and poor Americans are aware thatincome inequality in the United States has been rising since at least the 1960s.Furthermore, awareness of increasing income inequality is nontrivial: peopleare also concerned about increasing income inequality (Kelly and Enns,2010; McCall and Kenworthy, 2009). The public is also generally awarethat direct and indirect government expenditures have an effect on the econo-mic wellbeing of those in higher and lower income brackets and that peoplein those income brackets adjust their policy mood based on those governmentexpenditures (Ellis and Faricy, 2011; Kelly, 2009).

How and why public attitudes change due to increasing income ine-quality, however, has still not been fully explained. Kelly and Enns(2010) find that macro level public opinion changes as a result of increasingincome inequality. Their most significant discovery is that increasingincome inequality decreases public mood liberalism, and that this changehappens to those in the top and bottom income groups. Their conclusionappears to be counterintuitive yet is fully in line with Benabou’s conclusion(2000) that as long as government redistribution increases general welfareand income inequality is below a certain threshold, the public will prefermore conservative redistributive policies. However, their work examinesthe macro polity and does not attempt to discern what affects individualpolicy mood. Their only assumption about individual behaviour is that indi-viduals only care whether individual gains from redistribution are greaterthan the costs to their own financial situation.3

The Economic Roots of External Efficacy 5

Individual level research examining the effect of income inequality oncitizens’ views of politics emphasizes the possible negative consequences itmay have on representation of the middle class and the poor (Bartels, 2008;Gilens, 2009). Specifically, this line of individual, cross-sectional researchshows that the rich and the poor have differing opinions about how econo-mic resources should be distributed (Gilens, 2005; McCall and Kenworthy,2009). Despite these differences, it is very difficult to move policy awayfrom the status quo to change levels of inequality (Brady and Volden,1998; Enns et al., 2014; Gilens and Page, 2014; Krehbiel, 1998). If it isindeed moved, the evidence shows that policy makers are more likely tolisten to those at the top of the income scale than those at the bottom(Bartels, 2008; Garand, 2010; Gilens, 2005, 2009; Gilens and Page,2014; Schattschneider, 1960; Solt, 2010).

The ultimate conclusion of supporters of the unequal representationthesis is that as income inequality increases, those at the bottom of theincome distribution will forego participating in politics. E.E.Schattschneider (1960) first articulated the hypothesis that increasedincome inequality would lead to greater income bias and less participationgenerally because those with more economic resources would be able tocrowd off the political agenda policy alternatives favourable to themiddle class and the poor in favour of their own preferred policies. Solt(2010) finds evidence supporting this hypothesis by showing that voterturnout in gubernatorial elections declines in states with greater levels ofincome inequality. His analysis also shows that as the number of middleclass and lower class voters declines, the electorate in a state becomesbiased more toward its wealthiest voters.

Solt does not expressly test if increasing income inequality leads tolower levels of external political efficacy. However, an extension of hislogic creates the possibility that there may be a link between the twoconcepts. By definition, external political efficacy is related to the perceivedresponsiveness of government to the demands of its citizens (Chamberlain,2012). If Schattschneider’s hypothesis is correct and those with more eco-nomic resources are indeed able to remove policy alternatives beneficial tothe lower and middle classes from the political agenda, then it makes sensethat those in the lower and middle classes will begin to believe that theirgovernment is no longer responsive to their political demands. In short, itis likely that increasing income inequality may be an important factor inexplaining decreasing external political efficacy.

Linking Income Inequality and External Political Efficacy

I posit in this section that increasing income inequality in the United Statesis an important predictor of decreasing external political efficacy in the

6 MIKEL NORRIS

American electorate. There is considerable circumstantial evidence linkinggreater income inequality to declining external political efficacy. However,the inferred linkage has not been adequately explained or tested.

Theoretically, this analysis falls in line with the literature on unequalrepresentation arising from growing income inequality. At the heart ofthis perspective is the contention that economic inequality begets politicalinequality. This creates a “vicious cycle” where income inequalitybecomes self-reinforcing (Kelly and Enns, 2010: 857). The unequal repre-sentation perspective suggests that the reason income inequality becomesself-reinforcing is because the political inequalities arising from increasingincome inequality prevent the middle class and the poor from using govern-ment to curb economic disparities. The most obvious way the unequalrepresentation perspective can manifest itself is by observing the adoptionof policies that benefit the wealthy. To this end, Bartels (2008) finds that theSenate’s voting behaviour on roll call votes aligns more closely with thewealthy than with the poor. Gilens (2005) finds that, although the preferen-ces of the rich and poor are often congruent, policy makers are much morelikely to change policy to favour the rich when the policy preferences of therich and the poor diverge. The second way economic inequality can becomeself-reinforcing is via status quo bias that creates what Hacker and Pierson(2010) call policy drift. According to Hacker and Pierson, policy driftoccurs when government fails to adopt new policies to address new econo-mic realities.

The conclusion of this literature is that the wealthy are representedmore extensively in the political process. This comports with Gilens andPage’s evidence (2014) that the American political system operates infavour of economic elites and groups representing business interests.They find no evidence that the political system is responsive to average citi-zens or majoritarian interest groups. Their conclusions comport with theproposition that decreasing external political efficacy is fueled bygrowing income inequality. While measured at the individual level, thereare group and system-level components to external political efficacy.External political efficacy is affected by factors outside the individual andis also shaped by community or group membership. especially when agroup is homogenous (a racial minority, for example), speaks in acommon voice, or can demonstrate strength in numbers (Anderson, 2010;Koch, 1993; Mangum, 2003; Michelson, 2000). When citizens recognizethat government leaders and institutions are not responsive to the needsof their group but are responsive to the political needs of other groups,levels of external political efficacy should decline.

This also aligns with E.E. Schattschneider’s hypothesis that the sourceof suppression of the value of one’s vote is the crowding out of issuesdesired by the middle class and the poor, thereby narrowing the scope ofconflict (Schattschneider, 1960). Recall that Solt (2010) finds that


increasing income inequality at the state level depresses voter turnout ingubernatorial elections. What causes citizens to conclude that the valueof their vote has declined? Declining external political efficacy providesa sensible answer to this question. As income inequality increases, thewealthy will be better able to define the policy alternatives available tothe electorate.4 These alternatives, however, are not the desired alternativesof poor and middle class voters, and the increasing unavailability of resour-ces among these voters prevents them from defining policy alternatives inline with their interests.5 Furthermore, these voters also witness governmentrespond to the demands of wealthier groups in the political process.6 Theincreased inability of these citizens to define policy alternatives, coupledwith the perception that government is only responsive to the demands ofwealthier citizens, leads to the belief that government is not able orwilling to meet their needs. This leads to a decline in external political effi-cacy. It is only after this decline that we will see a decline in voter turnoutamong the poor and middle classes. This explanation leads to the followinghypothesis: Electorates with greater income inequality will have lowerlevels of external political efficacy.7

An alternative hypothesis of the relationship between income inequa-lity and external political efficacy is that the relationship is spurious, andthat other intervening variables are responsible for external political efficacy’sdecline. It could also be plausible that income inequality’s effect on externalpolitical efficacy is sporadic and only explains external political efficacy’sdecline at different points in time. First, it should be stated that income ine-quality is not a panacea. Other factors are quite likely driving down levels ofexternal political efficacy as well. However, I believe income inequalitywill exert its own independent effect on external political efficacy.Second, the models included in this analysis account for the effects oftime in two ways. The first is by including a variable accounting forwhether respondents were questioned during a presidential election year.The second is by the nature of the models themselves. The multilevelmodels used here calculate separate error terms for each state and eachyear, thus accounting for year and state-specific effects and reducing therisk of omitted variable bias (Steenbergen and Jones, 2002).

Data and Method

I analyze the effects of income inequality on external political efficacy in across-sectional analysis. It is important to analyze differing levels of incomeinequality and external political efficacy in a variety of contexts. Therefore,I focus on individual-level analyses of income inequality’s effect on exter-nal political efficacy in different states during different years. Doing so pro-vides the greatest possible variation in the variables of interest because both

8 MIKEL NORRIS

income inequality and external political efficacy vary greatly across statesand over time (Chamberlain, 2012, 2013; Kelly and Witko, 2012; Solt,2010). Figure 1 displays the variation in income inequality and externalpolitical efficacy in four different states: Colorado, Florida, Iowa, andOhio. A review of Figure 1 shows that there can be a great deal of variationin income inequality and external political efficacy over time both withinand across states. Furthermore, Figure 1 shows that, at least in these fourstates, external political efficacy has declined over time, while income ine-quality has risen.

Analyzing income inequality at the state level also has other benefits.States now play a much greater role in defining the winners and losers ofdistributional outcomes, especially since the Republican takeover ofCongress in 1994 (Kelly and Witko, 2012). While citizens recognize thegrowth of income inequality at the national level, they are more likely toobserve its effects locally via social comparison (Killian et al., 2008;Newman et al., 2015; Panning, 1983; Tversky and Kahneman, 1991).Furthermore, there is research finding that state governments performbetter and are more responsive to their citizenries when they have greaterlevels of social capital (Knack, 2002). To meet this end, this study analyzesindividual levels of external political efficacy from 1976 to 2004. Most ofthe data used in this analysis come from the American National ElectionStudies Cumulative Data File. The data were collected during every presi-dential and mid-term election year except for 1986 and 2002.8

External political efficacy

The dependent variable in this analysis is an individual’s external politicalefficacy index score as measured in the American National Election StudyCumulative Data File. The American National Election Study produces theindex by combining individual answers to two different statements. First,respondents are asked to agree or disagree with the following statement:“I don’t think public officials care much what people like me think.” Thesecond statement respondents are asked to agree or disagree with is:“People like me don’t have any say in what the government does.”Respondents who agreed with both statements are considered to have lowexternal political efficacy and are assigned a value of 0. Those who dis-agreed with both statements are considered to have high external politicalefficacy and are assigned a value of 100. Those who agreed with one ofthe two statements are considered to have moderate external political effi-cacy and are assigned a value of 50. To facilitate analysis, I recoded theindex so that respondents with a score of 50 were coded as 1, and respon-dents with a score of 100 were coded as 2. The resulting variable is a cate-gorical index ranging from 0 to 2.9


FIGURE 1Mean Levels of External Political Efficacy and Income Inequality for Selected States

10M

IKELN

ORRIS

Income inequality

I measure state-level income inequality by utilizing a post-transfer Ginicoefficient for households in each state for each year. This coefficientwas calculated by Kelly and Witko (2012) for each state from 1976 to2006, and is based on income data gathered from the US Census Bureauand the Annual Social and Economic Supplement. The Gini coefficientmeasures the deviation of the actual distribution of income from a hypothe-tically equal distribution of income. A Gini coefficient of zero means thatincome is distributed equally across all households in a state, while aGini coefficient of one means that income is distributed to only a singlehousehold in the state’s population. While there is ample variation in thevalue of the Gini coefficients across states, the Gini coefficients for allstates increased in value over the time period under consideration.

Individual-level variables

I control for several individual-level variables in this analysis. Ordinalscales measuring respondents’ income levels and educational attainmentare included in the models presented here. The seven-point party identifica-tion variable is included. Measures of age and interest in politics are alsoincluded in the models. Dummy variables are used to control for respon-dents’ race and gender. A dummy variable is also included to identify res-pondents who were employed when they were surveyed. All informationfor these variables was gathered from the National Election StudyCumulative Data File.

State and national variables

Contextual variables must be included in cross-sectional analyses to insureaccurate inferences based on individual behaviour (Achen and Shively,2003). Therefore, several state and national variables are included in thisanalysis to account for political and economic context. First, I include adummy variable for respondents residing in southern states to account forregional differences in external political efficacy. At the national level, adummy variable is included to account for whether the respondent was sur-veyed during a presidential election year. I also include a variable measur-ing the distance between the median member of the majority party in theSenate and the filibuster pivot in the Senate to account for the effects ofpolarization on external political efficacy.10

Two state-level contextual variables are included in this analysis. First,a union density variable for each state in each year is added to the model, andis calculated as the proportion of the nonagricultural workforce representedby a union in each state.11 A variable measuring the one-year change in


gross state product at the time of the respondent’s interview is included inthe analysis. These data were gathered from the US Department ofCommerce, Bureau of Economic Analysis. The descriptive statistics forall of the variables used in this analysis are presented in Table 1, alongwith predictions as to the expected direction of influence each variablemay have on external political efficacy.

Model Specification

The data used here are multilevel in nature. Most of the independent varia-bles used measure individual characteristics. However, individual respon-dents are nested within different states, and the states are in turn nestedwithin different years. This means that the observations are not truly inde-pendent from one another. Since these clustered observations violate theassumption of independent error terms, the models presented in this analy-sis are multilevel models. Multilevel models ensure that the standard errorsfor contextual variables (variables that vary at the state and national level)are not underestimated by calculating different error terms for each state andeach year. These separate error terms also capture contextual effects that areoutside the model and help protect against omitted variable bias(Steenbergen and Jones, 2002). This type of analysis allows for betterapproximation of individual and contextual variables by allowing the slopes

TABLE 1Descriptive Statistics

Dependent Variable Mean St. Dev. Min Max Expected Sign

External Political Efficacy Index .968 .854 0 2 −Individual-Level Variables

Income 1.92 1.14 0 4 +Education 3.04 1.72 0 5 +Age 45.5 17.15 17 95 −Party Identification 2.7 .206 0 6 +Gender .552 .497 0 1 +Race .203 .403 0 1 −Interest in Politics 1.82 .979 0 3 +Currently Employed .623 .485 0 1 +State/National Variables

State-Level Gini Coefficient .412 .03 .348 .509 −Yearly Change in Gross State Product .083 .04 −.053 .284 +Union Density .18 .082 0 .387 +Southern State .31 .462 0 1 −Filibuster Pivot .005 .073 −.096 .158 −

Sources: NES Cumulative Data File, Bureau of Economic Analysis, and Author’s Data.

12 MIKEL NORRIS

and intercepts of the individual variables to vary by state and also by year.States also have their own slopes and intercepts. Multilevel modeling thereforeimproves the chances that inferences drawn from nested data (in this case,state data) are more accurate (Tolbert et al., 2009). Multilevel ordinal logisticalregression with mixed effects is used to estimate all models due to the catego-rical nature of the external political efficacy index. Given a three-level datastructure such as this one (respondents nested in different states, and statesnested within different years) a categorical variable c = 1, …, C ordered cate-gories of the dependent variable, the latent response strength of respondent j incluster i at time k{yijk} can be modeled as the following:

yijk ¼ x0ijkβþ uij þ ui þ εijkwhere xijk is the covariate vector, β are the unknown parameters of the varia-bles in the model, vi is the unknown random effect at the yearly level, vij is theunknown effect at the state level, and εijk is the error term. Given the aboveequation, we can determine the probability of a respondent falling into oneof the three categories of the external political efficacy index, conditionalon the random effects υij and υi via the following equation:

PrðYijk ¼ cj υij; viÞ ¼ Ψ yc � zijk� �� yc�1 � zijk

� ��

where zijk = x′ijk β + υij + υi andΨ{.} represents the cumulative density function

(Raman and Hedeker, 2005). Analyses were performed using the meologitfunction in Stata 13.

Results12

The results of various multilevel ordinal logistic regressions measuring theeffect of income inequality on external political efficacy are presented inTable 2. Model 1 displays the results of the full model, while Model 4 dis-plays results without including the variable measuring the distance of thefilibuster pivot from the median senator. Model 7 displays the results ofan analysis excluding the variable measuring residency in the South. Thismodel attempts to account for collinearity between the southern dummyvariable and the variable measuring state-level union density. Across allmodels, the chi square statistic is large and significant, allowing us toreject the hypothesis that all of the variables in the model are equal to zero.

Consistent with previous models of efficacy, more educated and weal-thier citizens are more externally efficacious. So are those respondents whohave a greater interest in politics. Respondents are more likely to havehigher levels of external political efficacy if they identify with theDemocratic party. Women are more externally efficacious than men, and


TABLE 2The Effects of Income Inequality on External Political Efficacy

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

IndependentVariables Estimate

(Std.Error) Estimate






(Std.Error)

Income Inequality −2.38c (.904) −2.03a (1.08) −3.64a (1.69) −1.23a (.303) −3.78a (.856) −7.22a (1.34) −4.11a (.704)Individual-LevelVariables

Income .137a (.016) .144a (.020) .128a (.024) .135a (.015) .145a (.020) .127a (.024) .138a (.016)Education .235a (.011) .231a (.014) .244a (.017) .237a (.010) .228a (.014) .239a (.017) .232a (.011)Age −.003a (.001) −.002 (.001) −.006a (.002) −.003a (.001) −.002 (.001) −.006a (.002) −.004a (.001)Party Identification .022a (.008) .035a (.010) −.001 (.012) .021a (.008) .035a (.010) −.002 (.013) .022a (.008)Gender .172a (.032) .171a (.041) .176a (.052) .177a (.032) .175a (.041) .175a (.052) .175a (.032)Race −.073c (.041) −.052 (.052) −.110c (.067) −.067c (.040) −.055 (.052) −.117c (.067) −.069c (.041)Currently Employed .069c (.037) .058 (.047) .085 (.058) .067c (.036) .057 (.047) .086 (.058) .066c (.036)Interest in Politics .279a (.017) .284a (.022) .277c (.027) .284a (.167) .286a (.022) .279a (.027) .285a (.017)State/NationalVariables

Δ Gross StateProduct

1.03c (.561) −.013 (.757) −.874 (.811) −.089 (.522) .704 (.712) −.407 (.764) .201 (.518)

Untion Density −.038 (.395) −.235 (.488) −.072 (.623) .078 (.349) .481 (.409) .987c (.554) .133 (.261)Southern State −.039 (.065) −.079 (.080) .042 (.104) −.033 (.056) .022 (.071) .216b (.093)Filibuster Pivot −1.40a (.440) −1.53a (.590) −5.41a (.722)Presidential ElectionYear

.251a (.041) .598a (.042) .406a (.040)

VarianceComponents

State-Level .253 (.029) .341 (.054) .497 (.084) .084 (.009) .324 (.052) .514 (.083) .217 (.024)State/Year Level .045 (.011) .039 (.013) .045 (.017) (.040) (.010) .041 (.013) .036 (.017) (.043) (.010)χ2 1,669.1a 999.42a 753.23a 1,822.94a 991.64a 654.32a 1,709.41a

14M

IKELN

ORRIS

Log-Likelihood −16419.96 −10122.31 −6281.72 −16419.96 −10125.64 −6282.79 −16418.52IndividualObservations

16,274 9,951 6,323 16,274 9,951 6,323 16,274

Years 13 8 5 13 8 5 13State/Years 452 284 168 452 284 168 452

a. p≤ .10; b. p≤ .05; c. p≤ .01.Source: American National Election Study, Cumulative Data File; U.S Bureau of Labor Statistics; Author’s Data Coefficients are ordinal logistic regression coeffi-ceints, calculated using meologit in Stata 13.

The

Econom

icRoots

ofExternal

Efficacy

15

minorities are less likely, but not necessarily significantly less likely, to bemore externally efficacious that Caucasians. Age is also negatively relatedto external political efficacy, but its effect is small, and it is not significantacross all of the models. Being employed also increases the probability onewill have higher levels of external political efficacy, but its significance isnot consistent across all models.

There is a great degree of variation in the effect state and national varia-bles have in each of the models. Few of the state and national variables aresignificant across the range of models in Table 2. The most significant pre-dictors of external political efficacy in the full model are change in grossstate product, polarization and whether the respondent’s external politicalefficacy was measured during a presidential election year. Given the signi-ficant differences in respondents’ external political efficacy between presi-dential and midterm election years (t =−10.17, p ≤ .01), Models 2 and 5model the effect of income inequality on external political efficacy in pre-sidential years only. Models 3 and 6 measure the effect of income inequalityon external political efficacy during midterm election years only. Incomeinequality remains a large, significant and negative predictor of externalpolitical efficacy in these models as well. An especially interestingfinding in these models is that income inequality proves to be the largestpredictor of external political efficacy in most of the models presentedhere. Only in Model 3—during midterm election years—do we find thatpolarization is a greater predictor of external political efficacy.

Since logit coefficients are not readily interpretable, I used the fullmodel presented in Model 1 to compute predicted probabilities of atypical respondent having high or low levels of external political efficacyover the range of observable values of state-level income inequality. Theprobabilities for Model 1 are presented in Figure 2.13

As Figure 2 demonstrates, external political efficacy levels fall preci-pitously as state-level income inequality increases. The probability of a res-pondent having the highest levels of external political efficacy in a statewith the lowest levels of income inequality is .34 [.31, .41] and the proba-bility of a respondent having low levels of external political efficacy is .36[.29, .39]. The probability of a respondent having high levels of externalpolitical efficacy falls to .17 [.14, .22] as state-level income inequalityrises to its maximum value, a decline of .17. Conversely, the probabilityof a respondent having low external political efficacy in states with thehighest income inequality rises to .58 [.52, .64], a .22 increase.

The online appendix presents several robustness checks to the modelspresented here. Specifically, I examine whether other factors like having aRepublican or Democratic governor or president effects external politicalefficacy, or whether declining external political efficacy is a result of a res-pondent’s political party not being in control of government when the res-pondent was polled. While there is evidence that having a Republican

16 MIKEL NORRIS

FIGURE 2Effect of Income Inequality on External Political Efficacy (Full Model)a.) High External Political Efficacy b.) Low External Political Efficacy


president in power increases external political efficacy, and that respondentsare more externally efficacious when their party is in power, these robust-ness checks do not change the effect of income inequality on external poli-tical efficacy. Income inequality remains a large and robust predictor acrossall specifications. Furthermore, comparisons made with two standard pre-dictors of external political efficacy, income and education, show thatincome inequality rivals these two factors in magnitude of effect on externalpolitical efficacy.

Discussion and Conclusion

The purpose of this paper is to explore the possible relationship betweenincome inequality and external political efficacy. Although it is well documen-ted that external political efficacy has greatly declined in the American electo-rate, scholars have had difficulty in identifying the causes of this decline.Whilethere are certainly multiple factors working together to decrease levels of exter-nal political efficacy in the American electorate, it is likely that economicfactors have an important role to play in external political efficacy’s decline.

In this paper I hypothesize that increasing income inequality is animportant factor in the decline of external political efficacy. The analysesperformed in this paper bear out this conclusion: greater inequality doeslead to decreased levels of external political efficacy. Increasing incomeinequality drives down external political efficacy regardless of model spe-cification. These results are consistent across models, with only increasedpolitical polarization rivaling income inequality as the best predictor ofdeclining external political efficacy in the American electorate.

The results of this paper align nicely with previous research that showsincreasing income inequality leads to less political participation and decreasesthe likelihood that those in the middle and lower income brackets will vote(Schattschneider, 1960; Solt, 2010). Admittedly, the results of this paper areexploratory. It could be quite possible, though not examined here, that thereason people residing in states where external political efficacy is low andincome inequality is high perceive that the political interests being served intheir state are the interests of those that the top of the income distribution. Itcould also be possible that decreasing external political efficacy leads thoseat the bottom of the income distribution to forego going to the polls.Although neither of these propositions is tested here, both warrant futureresearch to determine if these connections are accurate. There is a wealth ofresearch that explores how local perceptions of prosperity determine how citi-zens view politics, and ultimately determine whether a citizen will or will notvote (Killian et al., 2008). By combining this research we can develop agreater understanding of how income inequality—and economic inequalitymore generally—affects how citizens vote, or if they even vote at all.

18 MIKEL NORRIS

The results of this paper also have implications for future research onthe self-reinforcing link between income inequality and public opinion.Evidence suggests that the effects of income inequality on public opinionare indeed self-reinforcing (Kelly and Enns, 2010; Solt, 2010); however,there is conflict between these two conclusions. Solt believes the effectof income inequality on participation is self-reinforcing because as electo-rates become smaller and wealthier, government will reward these electora-tes with fewer redistributive policies. Kelly and Enns, however, find that atthe macro level individuals from all income brackets favour more conserva-tive distributive policies as income inequality increases. A deeper unders-tanding of income inequality’s effect on external political efficacy mayhelp to reconcile these results. It is possible that the reason those in thelower and middle classes favour more conservative policies as income ine-quality increases is because they see redistributive policies as being highlyineffective. However, the reason these policies are ineffective is becausesmaller, wealthier electorates are getting exactly what they voted for:weaker redistributive policies. In essence, individuals from differentclasses may be taking different roads to the same destination. Having losttheir sense of external political efficacy, lower and middle class votersmay be shunning what they believe to be unresponsive government solu-tions for more market-based solutions to their economic problems.However, it could also be that the market-based solutions they areturning to do not assist them either because the boundaries of these solu-tions are dictated by the smaller, wealthier electorates who vote for econo-mic policies that protect their economic interests or, at least, prevent policyfrom changing the status quo (Enns et al., 2014; Hacker and Pierson, 2010).Another fruitful avenue for future research is to examine whether decliningexternal political efficacy leads individuals to prefer more conservative,market-based policy outcomes from government.

Supplementary materials

To view Supplementary material for this article, please visit http:dx.doi.org/10.1017/S0008423915001080

Notes

1 The initial starting point for this debate begins with conclusions in the AmericanPolitical Science Association’s Taskforce on Inequality and American Democracy(2006). The taskforce concluded that increasing inequality has led to a decline in trustin government and an increased concern about government responsiveness. Bennett(2006) concludes the taskforce’s conclusions are tenuous, at best. However, Bennettis clear that the weak relationship is only present in the 2000 and 2004 NationalElection Studies. This analysis uses a more extensive time frame.


http:dx.doi.org/10.1017/S0008423915001080http:dx.doi.org/10.1017/S0008423915001080

2 Currently, respondents to the two questions discussed above are able to answer each sta-tement with “agree,” “disagree,” or “neither.” “Agree” responses are coded as 100, “dis-agree” responses are coded as 0 and “neither” responses are coded as 50. The answersare then indexed and aggregated into an external political efficacy index ranging from 0to 100. This variable can be used to measure external political efficacy in the entire elec-torate. An individual external political efficacy index ranging from 0 to 100 exists aswell. However, measurement of this variable was changed in 1988 to allow for morevariation. Scholars interested in analyzing external political efficacy before and after1988 commonly collapse the new and old measure into an ordered categorical index.This is the method for measuring external political efficacy used in this paper.

3 See Kelly and Enns (2010, fn 10) for a more specific discussion of the underlying pre-sumptions of Benabou’s theory that is tested in their paper.

4 There is a growing debate around this issue in the literature on political representationand campaign finance reform. Current research recognizes that the increasing abilityof wealthier citizens to contribute more resources to political campaigns raises concernsover political equality (see Grant and Randolph, 2003). Political decisions like theSupreme Court’s decision in Citizen’s United v. Federal Election Commission (2010)and the great influx of campaign contributions in the 2012 presidential election haveonly heightened the need to understand the effects of unequal campaign contributionson political representation and—more specifically related to this study—the ability toshape political agendas.

5 An interesting point raised by a reviewer is that the calculus of turnout and voting underthe conditions posited by Schattschneider, Solt and in this paper hinge on voters, andespecially lower-class voters, voting based on issue positions. However, it is wellknown that lower-class voters do not regularly vote on issue positions because theyare generally unknowledgeable about political issues. Although this is true, there is exi-sting literature that shows that lower-income voters decide not to vote based not on issuepositions but rather on the perception that their vote will do nothing to benefit them indi-vidually nor benefit the groups to which they belong. Griffin and Newman find someevidence of this effect by showing that, when seeking re-election, incumbent Membersof Congress forego pursuing votes from groups with low voting power. They also reco-gnize the possible effect this could have on external political efficacy. In a supplemen-tary analysis, they use the same components of their model on voting power and policyrepresentation to show that nonvoters are significantly less externally efficacious thanvoters. They readily admit that their analysis is preliminary (2012, fn 18); however, itlends support to the belief that while low and middle income voters may not be know-ledgeable about policy issues, they can at least perceive whether government makesdecisions that benefit them as a group.

6 Support for this claim can be derived from Gilens and Page (2014) who find evidencethat government is only responsive to the demands of business-oriented interest groupsand economic elites. While this study does not support the contention that these groupsare able to remove policies off the political agenda, it does lend support to the contentionthat middle and lower class voters perceive that government is unresponsive to theirinterests. It is possible that this perception leads these voters to believe policies benefi-cial to their economic well-being are not on the political agenda. Indeed, Gilens alludesto the possible negative effect biased policy responsiveness may have on external poli-tical efficacy when he stated: “Most middle-income Americans think that public officialsdo not care much about the preferences of ‘people like me.’ Sadly, the results presentedabove suggest they might be right” (Gilens, 2005: 794).

7 Future research may want to incorporate income inequality and external political effi-cacy together in a way to determine actual voter turnout or vote choice. These factorscould be important in estimating simultaneous equation models of voter turnout and

20 MIKEL NORRIS

choice, or models of voter turnout and choice over time. These models, however, arebeyond the scope of this paper. Here, I am merely attempting to explain how incomeinequality and external political efficacy may be related.

8 Data were not gathered for the statement “People like me don’t have any say about whatthe government does” (one component of the external political efficacy index) in 1986.Data on respondents’ income quintiles were not gathered in 2002.

9 Prior to 1988 respondents were only able to answer “agree” or “disagree” to the externalpolitical efficacy questions in the National Election Study. After 1988 respondents wereable to answer “neither agree nor disagree.” Answering in this way enabled respondentsafter 1988 to score 0, 25, 50, 75, and 100 on the external political efficacy index. Forstandardization purposes, respondents scoring 25 after 1988 were coded as0. Respondents scoring 75 after 1988 were coded as 1. See Chamberlain (2013) foran example of the external political efficacy index coded and used in this way.

10 Other measures of political polarization were included in previous models but were notused because if high collinearity between polarization and income inequality. This is notan uncommon problem in research on income inequality that accounts for polarization(see Enns et al., 2014; McCarty et al., 2006). Using the distance between the majorityparty median and the filibuster pivot in the Senate is an acceptable measure of polariza-tion and also has the benefit of acting as a measure of gridlock in the political system (seeEnns et al., 2014, and Primo et al., 2008, for discussions of this topic).

11 Correlation between union density and Southern residence proved to be quite high(r = −.69, p ≤ .01), and does affect some of the results of the models presented herein.Model 7 in Table 2 displays results that exclude the Southern residence variable.

12 Additional analyses and robustness checks are available in the online appendix.13 Predicted probabilities were generated using Clarify (King et al., 2000). All variables

were held at their means (or modes) to estimate the likelihood of having either highor low external political efficacy.

References

Abramson, Paul R. and John H. Aldrich. 1982. “The Decline of Electoral Participation inAmerica.” American Political Science Review 76: 502–21.

Achen, Christopher and William Shively. 2003. Cross-Level Inference. Chicago: Universityof Chicago Press.

Anderson, Mary R. 2010. “Community Psychology, Political Efficacy, and Trust.” PoliticalPsychology 31: 59–84.

APSA Task Force on Inequality and American Democracy. 2006. “American Democracy inan Age of Rising Inequality.” Perspectives on Politics 2: 651–66.

Balch, George I. 1974. “Multiple Indicators in Survey Research: The Concept ‘Sense ofPolitical Efficacy.’” Political Methodology 1: 1–43.

Bartels, Larry M. 2000. “Partisanship and Voting Behavior, 1952–1996.” American Journalof Political Science 44: 35–50.

Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of a New Gilded Age.Princeton: Princeton University Press.

Benabou, Roland. 2000. “Unequal Societies: Income Distribution and the Social Contract.”American Economic Review 90: 96–129.

Bennett, Stephen Earl. 2006. “Comment on ‘American Democracy in an Age of RisingInequality.’” PS: Political Science and Politics 39: 51–4.

Bowler, Shaun and Todd Donovan. 2002. “Democracy, Institutions, and Attitudes aboutCitizen Influence on Government.” British Journal of Political Science 32: 371–90.

Brady, David and Craig Volden. 1998. Revolving Gridlock: Politics and Policy from Carterto Clinton. Boulder: Westview Press.


Chamberlain, Adam. 2012. “A Time-Series Analysis of External Efficacy.” Public OpinionQuarterly 76: 117–30.

Chamberlain, Adam. 2013. “The (Dis)connection Between Political Culture and ExternalEfficacy.” American Politics Research 41: 761–82.

Craig, Stephen C. 1980. “The Mobilization of Political Discontent.” Political Behavior 2:189–209.

Citizens United v. Federal Election Commission, 558 U.S. 310, 130 S.Ct. 876.Craig, Stephen C., Richard G. Niemi and Glenn E. Silver. 1990. “Political Efficacy and Trust:

A Report on the NES Pilot Study Items.” Political Behavior 12: 289–314.Ellis, Christopher and Christopher Faricy. 2011. “Social Policy and Public Opinion: How the

Ideological Direction of Spending Influences Public Mood.” Journal of Politics 73:1095–1110.

Enns, Peter K., Nathan J. Kelly, Jana Morgan, Thomas Volscho, and Christopher Witko.2014. “Conditional Status Quo Bias and Top Income Shares: How U.S. PoliticalInstitutions Have Benefitted the Rich.” Journal of Politics 76: 289–303.

Finkel, Stephen E. 1985. “Reciprocal Effects of Participation and Political Efficacy: A PanelAnalysis.” American Journal of Political Science 29: 891–913.

Garand, James C. 2010. “Income Inequality, Party Polarization, and Roll-Call Voting in theU.S. Senate.” Journal of Politics 72: 1109–28.

Gilens, Martin. 2005. “Inequality and Democratic Responsiveness.” Public OpinionQuarterly. 69: 778–96.

Gilens, Martin. 2009. “Preference Gaps and Inequality in Representation.” PS: PoliticalScience and Politics 42: 335–41.

Gilens, Martin. 2012. Affluence and Influence: Economic Inequality and Political Power inAmerica. Princeton: Princeton University Press.

Gilens, Martin and Benjamin I. Page. 2014. “Testing Theories of American Politics: Elites,Interest Groups, and Average Citizens.” Perspectives on Politics 12: 564–81.

Grant, J. Tobin; and Thomas J. Rudolph. 2003. “Value Conflict, Group Affect, and the Issueof Campaign Finance.” American Journal of Political Science 47: 453–69.

Griffin, John D. and Brian Newman. 2012. “Voting Power, Policy Representation, andDisparities in Voting’s Rewards.” Journal of Politics 75: 52–64.

Hacker, Jacob S. and Paul Pierson. 2010. Winner-Take-All Politics. New York: Simon &Schuster.

Hetherington, Marc J. 1998. “The Political Relevance of Political Trust.” American PoliticalScience Review 92: 791–808.

Kelly, Nathan J. 2005. “Political Choice, Public Policy, and Distributional Outcomes.”American Journal of Political Science 49: 865–80.

Kelly, Nathan J. 2009. The Politics of Income Inequality in the United States. Cambridge:Cambridge University Press.

Kelly, Nathan J. and Peter K. Enns. 2010. “Inequality and the Dynamics of Public Opinion:The Self-Reinforcing Link between Inequality and Mass Preferences.” AmericanJournal of Political Science 54: 855–70.

Kelly, Nathan J. and Christopher Witko. 2012. “Federalism and American Inequality.”Journal of Politics 74: 414–26.

Killian, Mitchell, Ryan Schoen and Aaron Dusso. 2008. “Keeping up with the Joneses: TheInterplay of Personal and Collective Evaluations in Voter Turnout.” Political Behavior30: 323–40.

King, Gary, Michael Tomz and Jason Wittenberg. 2000. “Making the Most of StatisticalAnalyses: Improving Interpretation and Presentation.” American Journal of PoliticalScience. 44: 347–61.

Knack, Stephen. 2002. “Social Capital and the Quality of Government: Evidence from theStates.” American Journal of Political Science 46: 772–85.

22 MIKEL NORRIS

Koch, Jeffrey R. 1993. “Assessments of Group Influence, Subjective Political Competence,and Interest Group Membership.” Political Behavior 15: 309–25.

Krehbiel, Keith. 1998. Pivotal Politics: A Theory of U.S. Lawmaking. Chicago: University ofChicago Press.

Lewis-Beck, Michael S. and Mary Stegmaier. 2000. “Economic Determinants of ElectoralOutcomes.” Annual Review of Political Science 3: 183–220.

Mangum, Maurice. 2003. “Psychological Involvement and Black Voter Turnout.” PoliticalResearch Quarterly 56: 41–48.

McCall, Leslie and Lane Kenworthy. 2009. “Americans’ Social Policy Preferences in the Eraof Rising Inequality.” Perspectives on Politics 7: 459–84.

McCarty, Nolan, Keith T. Poole and Howard Rosenthal. 2006. Polarized America: TheDance of Ideology and Unequal Riches. Cambridge MA: MIT Press.

Michelson, Melissa R. 2000. “Political Efficacy and Electoral Participation of ChicagoLatinos.” Social Science Quarterly 81: 136–50.

Newman, Benjamin J., Christopher D. Johnston and Patrick L. Lown. 2015. “FalseConsciousness or Class Awareness? Local Income Inequality, Personal EconomicPosition, and Belief in American Meritocracy.” American Journal of Political Science59: 326–40.

Niemi, Richard G., Stephen C. Craig, and Franco Mattei. 1991. “Measuring Internal PoliticalEfficacy in the 1988 National Election Study.” American Political Science Review 85:1407–13.

Panning, William H. 1983. “Inequality, Social Comparison, and Relative Deprivation.”American Political Science Review 77: 323–29.

Platt, Matthew B. 2008. “Participation for What? A Policy-Motivated Approach to PoliticalActivism.” Political Behavior 30: 391–413.

Pollock, Philip H., III. 1983. “The Participatory Consequences of Internal and ExternalPolitical Efficacy: A Research Note.” The Western Political Quarterly 36: 400–09.

Primo, David M., Sara A. Binder and Forrest Maltzman. 2008. “Who Consents? CompetingPivots in Federal Judicial Selection.” American Journal of Political Science 52: 471–89.

Raman, Rema and Donald Hedeker. 2005. “A Mix-Effects Regression Model for Three-Level Ordinal Response Data.” Statistics in Medicine 24: 3331–45.

Schattschneider, E.E. 1960. The Semisovereign People: A Realist’s View of Democracy inAmerica. New York: Holt, Rinehart & Winston.

Schur, Lisa A. and Douglas L. Kruse. 2000. “What Determines Voter Turnout? Lessons fromCitizens with Disabilities.” Social Science Quarterly 81: 571–87.

Solt, Frederick. 2010. “Does Economic Inequality Depress Electoral Participation? Testingthe Schattschneider Hypothesis.” Political Behavior 32: 285–301.

Soroka, Stuart N. and Christopher Wlezien. 2008. “On the Limits of Inequality inRepresentation.” PS: Political Science and Politics. 41: 319–27.

Steenbergen, Marco R. and Bradford S. Jones. 2002. “Modeling Multilevel Data Structures.”American Journal of Political Science 46: 218–37.

Teixeira, Ruy A. 1987. Why Americans Don’t Vote: Turnout Decline in the United States,1960–1984. Westport: Greenwood Press.

Timpone, Richard J. 1998. “Structure, Behavior, and Voter Turnout in the United States.”American Political Science Review 92: 145–58.

Tolbert, Caroline J., Daniel C. Bowen and Todd Donovan. 2009. “Initiative Campaigns:Direct Democracy and Voter Mobilization.” American Politics Research 37: 155–92.

Tversky, Amos and Daniel Kahneman. 1991. “Loss Aversion in Riskless Choice: AReference-Dependent Model.” Quarterly Journal of Economics 106: 1039–61.

Valentino, Nicholas A. Krysha Gregorowicz, and Eric W. Groenendyk. 2009. “Efficacy,Emotions and the Habit of Participation.” Political Behavior 31: 307–30.


The Economic Roots of External Efficacy: Assessing the Relationship between External Political Efficacy and Income InequalityIntroductionExternal Political Efficacy: Definition, Measurement and DeclineExternal Political Efficacy and the Economy: The Role of Income InequalityLinking Income Inequality and External Political EfficacyData and MethodExternal political efficacyIncome inequalityIndividual-level variablesState and national variables

Model SpecificationResults12Discussion and ConclusionSupplementary materialsReferences

the economic roots of external efficacy: assessing the … · the economic roots of external...

Documents